Research on abnormal traffic classification and detection technology based on machine learning

• 1、School of Electronic and Information Engineering,Liaoning Technical University,Huludao Liaoning 125105

Abstract：With the rapid development of emerging technologies such as 5G, the Internet of Things and blockchain, the scale and number of applications of the network are growing exponentially, dramatically changing people\'s lives. However, this is accompanied by increasingly complex network security threats, and the limitations of traditional network architectures lead to challenges in network management and security maintenance. In this context, this research focuses on the identification and classification of network traffic and abnormal traffic detection technology, proposes a network traffic identification method combining deep packet detection and machine learning to ensure the classification accuracy and real-time, and develops an abnormal traffic detection algorithm based on the combination of 1D CNN and LSTM. It can learn the spatial and temporal characteristics of network data at the same time, giving full play to the ability of CNN to learn the spatial characteristics of data and the ability of LSTM to learn the temporal characteristics of data. The test of this algorithm on the NSL-KDD dataset shows that its accuracy rate reaches 98.052%, which is about 15% and 10% higher than that of only using CNN or LSTM, respectively, and effectively improves the performance of network traffic management and anomaly detection.

Keywords： Traffic classification Deep packet inspection Machine learning Abnormal traffic detection algorithm